Exploiting Sparsity in the Difference Images to Achieve Higher Acceleration Factors in Non-Contrast MRA
Pippa Storey¹, Ricardo Otazo¹, Lazar Fleysher¹, Niels Oesingmann², Ruth P. Lim¹, Vivian S. Lee¹, Daniel K. Sodickson^1
1Radiology Department, NYU School of Medicine, New York, NY, United States; ²Siemens Medical Solutions USA

Non-contrast techniques for peripheral MRA exploit the pulsatility of arterial blood flow and involve subtraction of dark-blood images, acquired during fast flow, from bright-blood images, acquired during slow flow. The difference images, which depict the arteries, are sparse, although the source images are not. We show that higher acceleration factors can be achieved by performing subtraction on the raw data, before calculation of the GRAPPA weights, rather than on the final magnitude images. Depiction of large arteries is similar to that achieved with low acceleration factors and standard reconstruction, but depiction of small arteries and fine branch vessels is compromised.

Combination of Compressed Sensing and Parallel Imaging for Highly-Accelerated 3D First-Pass Cardiac Perfusion MRI
Ricardo Otazo¹, Jian Xu^2,3, Daniel Kim¹, Leon Axel¹, Daniel K. Sodickson^1
1Center for Biomedical Imaging, New York University School of Medicine, New York, NY, United States; ²Siemens Medical Solutions USA, New York, NY, United States; ³Polytechnic Institute of NYU, Brooklyn, NY, United States

Compressed sensing and parallel imaging are combined into a single joint acceleration approach for highly accelerated 3D first-pass cardiac perfusion MRI. 3D perfusion imaging is a natural candidate for this combined approach, due to increased sparsity and incoherence provided by the high dimensionality of the data, multi-dimensional acceleration capability and increased baseline SNR. We demonstrate the feasibility of high in vivo acceleration factors of 16 for 3D first-pass cardiac perfusion MRI studies with whole-heart coverage per heartbeat using a 32-element coil array

Efficient L1SPIRiT Reconstruction (ESPIRiT) for Highly Accelerated 3D Volumetric MRI with Parallel Imaging and Compressed Sensing
Peng Lai¹, Michael Lustig^2,3, Anja CS. Brau¹, Shreyas Vasanawala⁴, Philip J. Beatty¹, Marcus Alley^2
1Applied Science Laboratory, GE Healthcare, Menlo Park, CA, United States; ²Electrical Engineering, Stanford University, Stanford, CA, United States; ³Electrical Engineering and Computer Science, University of California, Berkeley, CA, United States; ⁴Radiology, Stanford University, Stanford, CA, United States

Conventional L1SPIRiT reconstruction enables highly-accelerated MRI by combining parallel imaging and compressed sensing but suffers from impractically long reconstruction time. This work developed a new efficient L1SPIRiT algorithm (ESPIRiT) to address the computation challenge from three perspectives: 1. reducing the computation complexity based on Eigenvector calculations, 2. reducing the number of pixels to process based on pixel-specific convergence, 3. reducing the number of iterations using parallel imaging initialization. ESPIRiT was compared with L1SPIRiT on in-vivo datasets. Our results show that ESPIRiT can improve image quality and reconstruction accuracy with >10× faster computation compared to L1SPIRiT.

Accelerated 3D Phase-Contrast Imaging Using Adaptive Compressed Sensing with No Free Parameters
Kedar Khare¹, Christopher J. Hardy¹, Kevin F. King², Patrick A. Turski³, Luca Marinelli^1
1GE Global Research Center, Niskayuna, NY, United States; ²GE Healthcare, Waukesha, WI, United States; ³School of Medicine & Public Health, University of Wisconsin, Madison, WI, United States

We present a robust method for compressed-sensing reconstruction using a data-driven, iterative soft-thresholding (ST) framework with no tuning of free parameters. The algorithm combines a Nesterov-type optimal gradient scheme for iterative update with adaptive wavelet denoising methods. Vascular 3D phase-contrast scans on healthy volunteers are used to show that image quality is comparable to that of empirically tuned, nonlinear conjugate-gradient (NLCG) reconstruction. Statistical analysis of image quality scores for five datasets indicates that the ST approach improves the robustness of the reconstruction and image quality as compared to NLCG with a single set of tuning parameters for all scans.

Nonconvex Compressive Sensing with Parallel Imaging for Highly Accelerated 4D CE-MRA
Joshua D. Trzasko¹, Clifton R. Haider¹, Eric A. Borisch¹, Stephen J. Riederer¹, Armando Manduca^1
1Mayo Clinic, Rochester, MN, United States

CAPR is a state-of-the-art Cartesian acquisition paradigm for time-resolved 3D contrast-enhanced MR angiography that typically employs Tikhonov and partial Fourier methods for image reconstruction.  When operating at extreme acceleration rates, such reconstructions can exhibit significant noise amplification and Gibbs artifacts, potentially inhibiting diagnosis.  In this work, an offline reconstruction framework for both view-shared and non-view-shared CAPR time-series acquisitions based on nonconvex Compressive Sensing is proposed and demonstrated to both suppress noise amplification and improve vessel conspicuity.

Fast MR Parameter Mapping from Highly Undersampled Data by Direct Reconstruction of Principal Component Coefficient Maps Using Compressed Sensing
Chuan Huang¹, Christian Graff², Ali Bilgin³, Maria I. Altbach^4
1Mathematics, University of Arizona, Tucson, AZ, United States; ²Program in Applied Mathematics, University of Arizona, Tucson, AZ, United States; ³Biomedical Engineering, University of Arizona, Tucson, AZ, United States; ⁴Radiology, University of Arizona, Tucson, AZ, United States

There has been an increased interest in quantitative MR parameter mapping techniques which enable direct comparison of tissue-related values between different subjects and scans. However the lengthy acquisition times needed by conventional parameter mapping methods limit their use in the clinic. In this work, we introduce a new model-based approach to reconstruct accurate T2 maps directly from highly undersampled FSE data. The proposed approach referred to as DIrect REconstruction of Principal COmponent coefficient Maps (DIREPCOM) removes non-linearity from the model and employs sparsity constraints in both the spatial and temporal dimensions to produces accurate T2 maps by using Principal Component Analysis. While this proposed technique has been illustrated for T2 estimation, the methodology can be adapted to the estimation of other MR parameters.

Compressed Sensing with Transform Domain Dependencies for Coronary MRI
Mehmet Akçakaya^1,2, Seunghoon Nam^1,2, Peng Hu², Warren Manning², Vahid Tarokh¹, Reza Nezafat^2
1Harvard University, Cambridge, MA, United States; ²Beth Israel Deaconess Medical Center, Harvard Medical School, Boston, MA, United States

Lengthy acquisition time is one of the main limitations of coronary MRI. Parallel imaging has been used to accelerate image acquisition but with limited success due to low signal-to-noise ratio. Compressed sensing (CS) has been recently utilized to accelerate image acquisition in MRI, but its use in cardiac MRI has been limited due to blurring artifacts. In this study, we develop an improved CS reconstruction method that uses the dependencies of transform domain coefficients to reduce the observed blurring and reconstruction artifacts in coronary MRI.

A Novel Approach for T1 Relaxometry Using Constrained Reconstruction in Parametric Dimension
Julia Velikina¹, Andrew L. Alexander¹, Alexey A. Samsonov^1
1University of Wisconsin - Madison, Madison, WI, United States

A novel method for T1 relaxometry is proposed using constrained reconstruction in the parametric (flip angle) dimension. Preliminary results indicate that the proposed method allows T1 estimation from undersampled data collected for multiple flip angles with  better accuracy than from the data collected for two ideal angles acquired within the same scan time.

Accelerated Breath-Hold Multi-Echo FSE Pulse Sequence Using Compressed Sensing and Parallel Imaging for T2 Measurement in the Heart
Li Feng¹, Ricardo Otazo², Jens Jensen², Daniel K. Sodickson², Daniel Kim^2
1Sackler Institute of Graduate Biomedical Sciences, New York University School of Medicine, New York, NY, United States; ²Radiology, New York University School of Medicine, New York, NY, United States

T2 Measurement can be used to detect pathological changes in tissue for a variety of clinical applications, including identification of edema and iron overload. Rapid T2 mapping in the heart is challenging because of the need to acquire adequate spatial resolution within clinically acceptable breath-hold duration of 20s or less. We propose to extend a recently developed breath-hold T2 mapping pulse sequence to achieve higher spatial resolution, by implementing a joint reconstruction algorithm that combines compressed sensing and parallel imaging. This accelerated T2 mapping pulse sequence with high spatial resolution was validated in vitro and in vivo.

Interleaved Variable Density Sampling with ARC Parallel Imaging and Cartesian HYPR for Dynamic MR Angiography
Kang Wang¹, James Holmes², Reed Busse², Philip Beatty³, Jean Brittain², Christopher Francois⁴, Lauren Keith¹, Yijing Wu¹, Frank Korosec^1,4
1Medical Physics, University of Wisconsin-Madison, Madison, WI, United States; ²Applied Science Laboratory, GE Healthcare, Madison, WI, United States; ³Applied Science Laboratory, GE Healthcare, Menlo Park, CA, United States; ⁴Radiology, University of Wisconsin-Madison, Madison, WI, United States

Both high spatial and temporal resolution are desired for contrast-enhanced MR angiography (CE-MRA). In this work, we describe a technique that combines interleaved variable density (IVD) Cartesian sampling, ARC parallel imaging (PI), and Cartesian HYPR reconstruction.  This technique is validated in multiple exams performed on healthy volunteers.